27 October 2017
The recent House of Lords Science and Technology Committee evidence session Can the Life Sciences Industrial Strategy become a reality? provides another timely opportunity to reflect upon the Life Sciences Industrial Strategy.
Previous PHG blogs have covered both the Life Sciences Industrial Strategy in general, finding the strategy to be an important step forward for the sector, and the regulatory proposals in particular, finding this particular section lacking in detail. However, arguably the most significant and timely part of the report concerns data. This section of the document provides key recommendations that will probably advance the interoperability of data, helping ensure the UK reaps the full benefits of the data revolution, including for health.
Past PHG blogs and reports have spoken at length regarding the factors that limit the sharing of genomic data. This body of work points to common barriers, and also common benefits to these barriers being disassembled.
There are clear and immediate advantages to unlocking data. Indeed, the Chief Medical Officer (CMO) Prof Dame Sally Davis makes clear in her Generation Genome report that reforms to data protection and the methods used to obtain consent are crucial if genomic medicine is to advance. Rich, interoperable datasets are essential both for high-quality research and for translating this research into better clinical services and outcomes for patients. As the CMO stresses, if we rely solely on research published in traditional journals, we act on evidence that already 12-24 months old. Given the pace of genomics and personalised medicine in general, it is clear that this time lag must be shortened. So, why the lack of progress?
In 2015, PHG and the Association of Clinical Genomic Science (ACGS) held a joint workshop focused on data sharing. This workshop highlighted the challenges to sharing data, the three most significant being:
While the Industrial Strategy is primarily a strategy for industry, NHS collaboration is nevertheless essential for the sector. Moreover, many of datasets industry requires will be the same or face similar challenges to their sharing. In this way, data sharing ought to be a core concern for the industry and is reflected as such in the report.
While the Industrial Strategy is primarily a strategy for industry, NHS collaboration is nevertheless essential
The report stresses that the UK is home to world-leading data sources in the form of Clinical Practice Research Datalink (CPRD), UK Biobank, and Genomics England. Respectively, each will hold information on 22 million patients, 500,000 participants, and up to 100,000 human genomes. Yet, the report also stresses that these data sources are only linked in some cases and do not currently reach their potential in providing ‘deep, real-time’ data for research. While these datasets are by no means the only rich pools of data in the UK, they do represent the issue in microcosm: rich data often exists but it is the (lack of) linkage and access that halts progress.
The Industrial Strategy proposes some solutions to these linking and access issues. The report stresses the need for national leadership to establish proper infrastructure and sharing mechanisms. In short, the report has two main, practical suggestions:
In regards to DIHs, the report cites NHS England’s Digital Maturity Index. This index makes plain that digital maturity and readiness for real-world evidence studies varies widely from region to region in the UK. What is needed then is a solution that both unifies and advances digital maturity nation-wide. As an answer, the report recommends the creation of 2-5 of DIHs across the country. These hubs are envisioned to secure data for primary, secondary, tertiary, social, and community care. In addition to establishing these hubs, the report also proposes that national standards of interoperability be set by NHS Digital. Effectively, this would allow studies to be run across national rather than just regional datasets. In short, these recommendations potentially answer one of the main barriers to data sharing: patchy, disconnected infrastructure.
Rare and orphan diseases require particularly large datasets supported by epidemiological findings if research is to progress, and will hence benefit from national, unified registries. Given this, further national registries are sorely needed.
In regards to national registries, the report also recognises the success of national data registries for certain therapeutic areas, citing the example of mental health registries in the form of the Clinical Record Interactive Search system. Yet the Industrial Strategy also points out that other therapy areas might equally benefit from similar registries. In particular, rare and orphan diseases require particularly large datasets supported by epidemiological findings if research is to progress, and will hence benefit from national, unified registries. Given this, further national registries are sorely needed. The Industrial Strategy’s approach here is therefore timely and the suggestion that these be co-ordinated by relevant patient charities is welcome provided that sufficient funding is available for this purpose.
The Industrial Strategy identifies regulatory heterogeneity and uncertainty as a major barrier to data sharing. Specifically, the report acknowledges that data sharing agreements and requirements vary between NHS trusts and other sources. The report proposes that national standards (following the recommendations of the National Data Guardian) ought to be developed to provide common agreements between trusts. Further, the report also identifies current linkage and access agreements to national databases as problematically long and unwieldy. Given this, what currently exists at a national level is need of streamlining. Accordingly, the report offers a sober view of the regulation that governs data sharing at the NHS trust level. In addition to this, the Industrial Strategy also offers pragmatic solutions to ensure that regulation facilitates rather than blocks prudent data sharing, and these are well-supported (for example, by the National Data Guardian).
The report offers sound advice for regulation at a trust level. However, it ought to be noted that regulation at this level is entirely contingent upon the UK and the EU’s legal position. The success or failure of these new data sharing arrangements therefore hinges upon the General Data Protection Regulation (GDPR) and the UK’s forthcoming Data Protection Bill. If these key pieces of legislation do not provide for fair, efficient data sharing, then effort spent on infrastructure and supervisory body regulation are undermined. In this way, the Industrial Strategy fails to provide a vision for what is necessary at the EU and national level to make sure regulation works for data instead of against it. Accordingly, the Industrial Strategy is bold and timely in most respects, but is left incomplete.
There is much in the Industrial Strategy to like. Above all, it is a strategy that has its priorities right: data sharing is a core concern and the report spends time outlining proportionate, welcome solutions to break data silos. Still, it ought to be noted that national and EU law has the power to scupper these best-laid plans. Hence, it is EU and national law that can act as the ultimate roadblock and should form the frame of any industrial plan.